Sorting device, display device, and sorting method

Abstract

To provide a classifying device, a display device, a printing device and a classifying method capable of identifying the visual characteristics of a user and of utilizing the identified visual characteristics.SOLUTION: A classifying device 1 includes a selecting unit 3 and a classifying unit 4. The selecting unit 3 allows a user to select a converted image CP that is the most similar to an appearance in actually viewing an object in an original image OP among multiple converted images CP obtained by performing image quality conversion on the original image OP at multiple tone curves TC-1 to TC-5 which have different conversion characteristics of brightness and of contrast. The classifying unit 4 classifies the user based on the selection result for the converted image CP by the user with the selecting unit 3 in accordance with a classification reference R generated by cluster analysis using the selection results of the converted images CP by multiple examinees.SELECTED DRAWING: Figure 6

Description

[Technical field]

[0001] The present invention relates to a classification device, a display device, reference and classification methods. [Background technology]

[0002] Image quality is adjusted so that no black or white areas occur in a displayed image (for example, Patent Document 1). [Prior art documents] [Patent documents]

[0003] [Patent Document 1] International Publication No. 2018 / 212212 Summary of the Invention [Problem to be solved by the invention]

[0004] Even when an image is adjusted as described above, different people will have different impressions when viewing the image. An image that does not look strange to one person may often look strange to another. To eliminate such strangeness, it is necessary to adjust the image quality of the displayed image to match the visual characteristics of each individual. However, adjusting image quality to match the visual characteristics of each individual user is necessarily a cumbersome task and is not realistic. At present, the reality is that no method has been found to identify the visual characteristics of an individual user.

[0005] The present invention has been made under the above circumstances, and provides a classification device, a display device, and a method for identifying a visual characteristic of a user and utilizing the identified visual characteristic. reference The purpose of this document is to provide a classification method for the above-mentioned substances. [Means for solving the problem]

[0006] In order to achieve the above object, a classification device according to a first aspect of the present invention comprises: a selection unit that allows a user to select a converted image that is closest to how a subject in an original image looks when actually viewed from among a plurality of converted images obtained by converting the image quality of an original image using a plurality of tone curves having different brightness and contrast conversion characteristics; a classification unit that classifies users based on the selection results of the converted images by the users in the selection unit according to classification criteria generated by cluster analysis using the selection results of the converted images by a plurality of subjects; Equipped with 、 In the cluster analysis, the classification criteria are generated by classifying a plurality of subjects in a vector space in which a coordinate axis indicating a numerical value corresponding to a selection result of a converted image that is closest to how the subject looks when actually viewed is independently defined for each expected time period when the subject is expected to be viewed; The selection unit is having a user select a converted image that is closest to how the subject appears when actually viewed for each of the estimated time periods, and determining a vector that indicates the user's selection result in the vector space; The classification unit includes: Users are classified based on vectors that indicate user selections in the vector space.

[0010] The present invention 2 The display device according to the aspect of the present invention comprises: First aspect of the present invention To the point Such a classification device, an image quality conversion unit that converts the image quality of a displayed image using a tone curve corresponding to a cluster of users classified by the classification device, among a plurality of tone curves having different brightness and contrast conversion characteristics; Equipped with.

[0013] The present invention 3 The classification method according to the above viewpoint is as follows: A classification method performed by a classification device, comprising: a cluster analysis is performed using the results of selecting a converted image that is closest to how the subject in the original image looks when actually viewed from a plurality of converted images obtained by converting the image quality of the original image using a plurality of tone curves each having different brightness and contrast conversion characteristics, for each estimated time period during which the plurality of subjects are expected to view the subject; and a classification criterion is generated for classifying the plurality of subjects in a vector space in which coordinate axes indicating values ​​corresponding to the selection results of the converted image that is closest to how the subject looks when actually viewed are independently defined for each estimated time period; The above From among multiple converted images, select the one that is closest to how the subject in the original image actually looks. , for each of the aforementioned estimated time periods Let the user choose to obtain a vector indicating a user's selection result in the vector space; before Record According to the classification criteria, In the vector space The users are classified based on their selection of transformed images. Effect of the Invention

[0014] According to the present invention, a cluster analysis is performed by having a plurality of subjects select a converted image that is closest to how a subject appears when actually viewed, and users are classified based on their selection results in accordance with classification criteria generated by the analysis, thereby identifying the user's visual characteristics and utilizing the identified visual characteristics. [Brief description of the drawings]

[0015] [Figure 1] 1 is a graph showing a tone curve. [Diagram 2] (A) is the original image for the morning, and (B) is the original image for the afternoon. [Diagram 3] FIG. 2 is a diagram showing an example of a converted image obtained from the tone curve of FIG. [Figure 4] 1 is a table showing the classification results of subjects. [Diagram 5] 1 is a graph showing changes in tone values ​​selected by subjects in the morning and afternoon. [Figure 6] FIG. 2 is a block diagram showing a software configuration of the classification device according to the first embodiment of the present invention. [Figure 7] FIG. 7 is a block diagram showing a hardware configuration of the classification device of FIG. 6. [Figure 8] 7 is a flowchart showing the operation of the classification device of FIG. 6. [Figure 9] 13 is a flowchart showing another example of a sequence for classifying users. [Figure 10] FIG. 7 is a block diagram showing a modified example of the software configuration of the classification device of FIG. [Figure 11] FIG. 11 is a block diagram showing a software configuration of a display device according to a second embodiment of the present invention. [Figure 12] FIG. 12 is a block diagram showing a configuration in which the display devices in FIG. 11 are connected via a communication network. [Figure 13] FIG. 12 is a block diagram showing a modified example of the software configuration of the display device in FIG. [Figure 14] FIG. 11 is a block diagram showing a software configuration of a printing device according to a third embodiment of the present invention. [Figure 15] 1 is a graph showing an example of a tone curve for printing. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0016] Hereinafter, an embodiment of the present invention will be described in detail with reference to the drawings. In each drawing, the same or equivalent parts are denoted by the same reference numerals.

[0017] <Embodiment 1> First, a first embodiment of the present invention will be described. A classification device 1 (see FIG. 6) according to this embodiment classifies users based on newly obtained knowledge about the visual characteristics of the users. First, a cluster analysis using the visual characteristics of the users used in the classification device 1 according to this embodiment will be described.

[0018] [Cluster analysis] Humans were subjected to 10 3.5 The natural impression felt when viewing a real landscape in an HDR environment can be perceived at a luminance ratio of about 10 2 Research is being conducted to establish a method for faithfully reproducing images and videos at a standard dynamic range (SDR) level perceptually. This research has revealed that in an HDR environment, humans do not directly perceive and remember the situation, but compress the image into an SDR level perceptual representation using global tone mapping corresponding to the time (morning, afternoon, night, etc.) and the environment (weather, indoors, outdoors, etc.). The compressed perceptual representation is called NVP (Normalized Visual Percept). It has been suggested that NVP may actively vary due to circadian rhythm, i.e., diurnal variation. In this embodiment, a visual characteristic in which NVP varies depending on the circadian rhythm is utilized, and multiple subjects are classified by cluster analysis using morning and afternoon images taken under environmental conditions of sunny outdoors.

[0019] In this cluster analysis, as shown in FIG. 1, five tone curves TC-1 to TC-5 with different conversion characteristics for brightness and contrast are used. The ones shown in FIG. 1 are normalized tone curves, and are defined as tone curves that do not cause blackout or whiteout at SDR levels. If tone curve TC-1 is used, an image that is dark overall but has high contrast is obtained. This image corresponds to standard perception outdoors in the morning (when the lifestyle is morning-type), for example. If tone curve TC-5 is used, an image that is bright overall but has low contrast is obtained. This image corresponds to standard perception at night. Tone curves TC-2 to TC-4 are intended to generate images with intermediate tones between the image converted by tone curve TC-1 and the image converted by tone curve TC-5.

[0020] In the cluster analysis according to the present embodiment, five tone curves TC-1 to TC-5 are used to convert the image quality of three original images A1 to A3 shown in FIG. 2(A) and three original images P1 to P3 shown in FIG. 2(B) to generate five converted images each. The original images A1 to A3 shown in FIG. 2(A) are images taken in the morning and each has a different subject. The original images P1 to P3 shown in FIG. 2(B) are images taken in the afternoon and each has a different subject. Note that although the images are shown in grayscale in FIG. 2(A) and FIG. 2(B), they are actually color images. The same is true for FIG. 3.

[0021] The five images on the left side of Fig. 3 are converted images obtained by converting the image quality of the same original image A1 shown at the left end of Fig. 2(A) using tone curves TC-1 to TC-5. The five images on the right side of Fig. 3 are converted images obtained by converting the image quality of the same original image P1 shown at the left end of Fig. 2(B) using tone curves TC-1 to TC-5.

[0022] In the cluster analysis, a plurality of subjects are made to select from among these multiple converted images the converted image that is closest to the appearance when the subject is actually viewed in the assumed time period (morning, afternoon). For example, from a total of 15 converted images (including the five converted images on the left in FIG. 3) obtained from the three original images A1 to A3 in FIG. 2(A), the subjects are made to select the converted image that is closest to the appearance in the morning for each of the original images A1 to A3. Furthermore, from a total of 15 converted images (including the five converted images on the right in FIG. 3) obtained from the three original images P1 to P3 in FIG. 2(B), the subjects are made to select the converted image that is closest to the appearance in the afternoon for the original images P1 to P3. The morning is assumed to be around 9:00 a.m., and the afternoon is assumed to be around 1:00 p.m. It is preferable that the selection be made in the early afternoon, when it is empirically known that the visual characteristics of the subjects are stable. The images shown to the subjects may be displayed on the display 14 (see FIG. 7) or may be printed on the printer 22 (see FIG. 7). It should be noted that the image printed on the printer is a converted image that has been subjected to additional pre-printing correction according to time using the printing tone curves TC-a1 and TC-a2, an example of which is shown in Figure 15.

[0023] In this cluster analysis, a classification criterion is generated by classifying a plurality of subjects in a vector space in which a coordinate axis indicating a tone value corresponding to a selection result of a converted image that is closest to the appearance when the subject is actually seen is independently defined for each of the original images A1 to A3 and P1 to P3 having different subjects or expected time periods. For example, when a converted image of tone curve TC-1 is selected, the coordinate value (tone value) related to the selection is 1. Similarly, when a converted image of tone curve TC-2 is selected, the tone value is 2, when a converted image of tone curve TC-3 is selected, the tone value is 3, and similarly, when a converted image of tone curves TC-4 and TC-5 is selected, the tone values ​​are 4 and 5. Cluster analysis is performed using these tone values. In this cluster analysis, a plurality of subjects are classified in a vector space of vectors indicating the user's selection results, whose elements are three tone values ​​indicating the converted images selected from the original images A1 to A3 in the morning and three tone values ​​indicating the converted images selected from the original images P1 to P3 in the afternoon. Hereinafter, this vector is also called a selection result vector.

[0024] In this embodiment, a plurality of subjects are classified by performing hierarchical cluster analysis by Ward's method on the main attributes and judgment results of the subjects. In Ward's method, a selection result vector is obtained for each subject, and classification is performed based on the distance in the vector space of the selection result vector for each subject.

[0025] In fact, when hierarchical cluster analysis was performed using Ward's method on 29 subjects, the subjects were classified into C1 and C2, C1 was further classified into C1-1 and C1-2, C1-1 was further classified into C1-1-1 and C1-1-2, and C1-2 was further classified into C1-2-1 and C1-2-2. Figure 4 shows the results of investigating regression characteristics using three explanatory variables: the amount of tone value fluctuation (absolute value) in the morning and afternoon, the direction of fluctuation, and the average value of the tone value. Figure 5 shows the fluctuation of tone values ​​in the morning and afternoon for these five clusters: C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2.

[0026] (1) C1 and C2 have different average tones for morning and afternoon. C2 is a type that maintains a high tone value at all times (low contrast similar to night even during the day). If you select an image with high contrast, the tone value will be low. (2) The direction of tone value change from morning to afternoon is different between C1-1 and C1-2. C1-2 is classified as a standard morning type. (3) C1-2 is classified into C1-2-1 and C1-2-2 depending on the magnitude of the gradient of the change in tone value. If the gradient is large, it is classified into C1-2-1, and if it is small, it is classified into C1-2-2. (4) C1-1 is classified into C1-1-1 and C1-1-2 according to the average tone value. If the average tone value is large, it is classified into C1-1-1, and if it is small, it is classified into C1-1-2. C1-1-1 belongs to the night type (a type in which contrast is low in the morning and increases in the afternoon). The human subjects selected images with high tone values ​​in the morning and images with low tone values ​​in the afternoon. (5) Subjects classified as C1-1-2 chose images with small tone values ​​in both the morning and afternoon, maintaining high contrast.

[0027] It should be noted that these classification results are merely examples, and the resulting clusters may change significantly if the number of subjects is increased.

[0028] A classification device 1 that classifies users using the above analysis results will be described. The classification device 1 according to this embodiment will be described as classifying users into any of C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2. However, the clusters are determined by the above-mentioned cluster analysis by multiple subjects, and the user may be classified into more clusters. As shown in FIG. 6, the classification device 1 includes a storage unit 2, a selection unit 3, and a classification unit 4.

[0029] [Storage] The memory unit 2 stores the following information: (A) Classification criterion R: The analysis criterion R is a classification criterion generated by cluster analysis using the selection results of multiple converted images CP by multiple subjects. It is information that serves as a criterion for classifying a user into one of the clusters. For example, the central vector of each cluster is the classification criterion. This is because users can be classified based on the distance between the user's selection result vector and the central vector of each cluster. This classification criterion is found by the cluster analysis device 7, which is an information processing device that actually performed the cluster analysis, and is stored in the storage unit 2. (B) Original image OP: For example, images A1 to A3 and P1 to P3 shown in Fig. 2(A) and Fig. 2(B), and the same images as those used in the cluster analysis are used. That is, a plurality of original images OP having different subjects are stored in the memory unit 2. When the classification device 1 generates a converted image CP from the original image OP, the memory unit 2 stores the original image OP. When the converted image CP is stored in the memory unit 2, it is not necessary to store the original image OP in the memory unit 2. (C) Transformed images CP: A plurality of transformed images CP obtained by transforming the image quality of the original image OP with the tone curves TC-1 to TC-5. These are the same images as the transformed images CP used when the classification standard R was generated. When the original image OP and the tone curves TC-1 to TC-5 are stored, there is no need to store the transformed images CP in the memory unit 2. (D) Tone curve TC: For example, tone curves TC-1 to TC-5. The same tone curves TC-1 to TC-5 used when the classification criterion R was generated are stored. When a converted image CP is stored, the classification device 1 does not need to store the tone curves TC-1 to TC-5. However, since it is necessary to obtain tone values ​​for user classification, the storage unit 2 needs to store which converted image CP corresponds to which tone value (which tone curve TC-1 to TC-5 corresponds to). As shown in FIG. 6, a plurality of original images OP, converted images CP, and tone curves TC are stored in the storage unit 2, but in the following drawings, for simplicity, each of these is illustrated one by one.

[0030] [Selection] The selection unit 3 allows the user to select the converted image CP that is closest to the appearance of the subject in the original image OP when actually viewed from among the multiple converted images CP. The converted image CP may be generated from the original image OP using a tone curve TC. For this reason, the selection unit 3 has a function of reading multiple converted images CP from the storage unit 2, displaying and outputting the converted images CP, displaying or outputting a sound to the user who has viewed the display so as to select the converted image CP, and inputting information on the converted image CP selected by the user. In this embodiment, the selection unit 3 allows the user to select the converted image CP that is closest to the appearance of the subject when actually viewed for each original image OP having a different subject, and obtains a numerical value (tone value) for each coordinate axis in the above-mentioned vector space. In addition, the selection unit 3 allows the user to select the converted image CP that is closest to the appearance of the subject when actually viewed for each expected time period (morning, afternoon), and obtains a numerical value (tone value) for each coordinate axis in the vector space. In this way, the user's selection result vector is determined.

[0031] [Classification Department] The classification unit 4 classifies users based on the selection result of the user's converted image CP by the selection unit 3 in accordance with the classification criterion R stored in the storage unit 2. In this embodiment, the classification unit 4 classifies users based on a vector represented by the user's selection result in a vector space. For example, the classification unit 4 classifies users into one of clusters C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2. The classification unit 4 outputs the classification result.

[0032] [Hardware configuration] Fig. 7 shows a hardware configuration of the classification device 1. As shown in Fig. 7, the classification device 1 includes a CPU (Central Processing Unit) 10, a memory 11, an external memory 12, an operation unit 13, a display 14, a communication interface (I / F) 15, an input / output unit 16, a timer 17, a camera 18A, and a microphone 18B. The components of the classification device 1 are connected via an internal bus 20.

[0033] The CPU 10 is a processor (arithmetic device) that executes a software program (hereinafter, simply referred to as a "program"). A program 19 is loaded into the memory 11 from an external memory 12, and the CPU 10 executes the program 19 stored in the memory 11 to realize the operations of the selection unit 3 and the classification unit 4.

[0034] The memory 11 is, for example, a random access memory (RAM). The memory 11 stores a program 19 to be executed by the CPU 10, and also stores data required for the CPU 10 to execute the program 19.

[0035] The external memory 12 is, for example, a hard disk. The external memory 12 stores a program 19 executed by the CPU 10. The program 19 is also stored in a recording medium 21 such as a portable USB (Universal Serial Bus) memory. The program 19 transferred from the recording medium 21 is stored in the external memory 12. The same is true for various data such as an original image OP, a converted image CP, a tone curve TC, and a classification criterion R.

[0036] The operation unit 13 is a man-machine interface that can be operated by a user. Examples of the operation unit 13 include a keyboard, a mouse, and a touch panel. The CPU 10 operates according to an operation input of the operation unit 13. For example, information on a converted image CP selected by a user is input by the operation input of the operation unit 13.

[0037] The display 14 is an image display device that displays an image. The display 14 outputs an image signal output from the CPU 10. As a result, an image based on the image signal is displayed on the display 14. In the case of a touch panel, the operation unit 13 and the display 14 are integrated. For example, a plurality of converted images CP are displayed on the display 14.

[0038] The communication interface 15 is a communication interface for communicating with external devices. The communication interface 15 is an input / output interface for a communication network. To the communication interface 15, for example, a printer 22 is connected.

[0039] The input / output unit 16 is an input / output interface for the recording medium 21. The program 19 or the various data described above is input via the input / output unit 16 and stored in the external memory 12.

[0040] The timer 17 measures time. The CPU 10 can detect the current time zone, for example, whether it is currently morning or afternoon, based on the time measured by the timer 17. The classification device 1 also includes a camera 18A (imaging device) and a microphone 18B (audio input device).

[0041] Next, the operation of the classification device 1 in Fig. 6 and Fig. 7, that is, the overall processing, will be described. As shown in Fig. 8, first, the selection unit 3 causes the display 14 to display the converted images CP and also causes the display 14 to display a question (step S10). Here, a plurality of converted images CP are displayed, and a question is displayed prompting the user to select one converted image CP for each time period and each subject that is closest to how the subject looks when actually viewed.

[0042] Next, the selection unit 3 waits until a selection input is made to the operation unit 13 (step S11; No). After viewing the multiple displayed converted images CP, the user selects a converted image CP that is similar to how the user sees it by operating the operation unit 13.

[0043] The above steps S10 and S11 correspond to a selection operation for having the user select the converted image CP closest to the appearance when the subject is actually viewed from among a plurality of converted images CP obtained by converting the image quality of the original image OP using a plurality of tone curves TC each having different conversion characteristics for brightness and contrast. Note that the selection unit 3 may operate in such a way that the morning converted image CP is displayed first to allow the user to select the converted image CP closest to the morning appearance first, and then the afternoon converted image CP is displayed to allow the user to select the converted image CP closest to the afternoon appearance later.

[0044] When a selection input is made (step S11; Yes), the classification unit 4 performs a classification process to classify the user based on the user's selection result by the selection unit 3 according to the classification criterion R stored in the storage unit 2 (step S12). Specifically, the classification unit 4 classifies the user based on which cluster in the vector space the selection result vector selected by the user in the selection unit 3 belongs to in light of the classification criterion R. For example, the clusters C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2 that are closest in the vector space to the user's selection result vector are determined to be the clusters to which the user belongs. After the classification process, the overall process ends.

[0045] In addition, when classifying into clusters C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2, the classification unit 4 may execute the process shown in Figure 9 and classify users based on the average tone value of the converted image CP selected by the selection unit 3, the direction of change in tone value due to the expected time period of the converted image CP selected by the selection unit 3, and the magnitude of the absolute value of the change.

[0046] (1) As shown in Fig. 9, the classification unit 4 calculates the average value of the tone values ​​of the converted images CP selected by the user in the morning and afternoon, and judges whether the calculated average value exceeds a first threshold value T1 (step S1). The first threshold value T1 is a value obtained by the above analysis. If the user is judged to be C2 (step S1; Yes), the classification unit 4 classifies the user as C2.

[0047] (2) If it is determined that the user belongs to C1 (step S1; No), the classification unit 4 determines whether the gradient of the change in tone value from morning to afternoon is positive (step S2). If it is determined that the gradient is positive (step S2; Yes), the classification unit 4 classifies the user as C1-2, and if it is determined that the gradient is not positive (step S2; No), the classification unit 4 classifies the user as C1-1.

[0048] (3) If the user is determined to be C1-1 (step S2; No), the classification unit 4 determines whether or not the average tone value exceeds the second threshold T2 (step S3). If the average tone value exceeds the second threshold T2 (step S3; Yes), the classification unit 4 classifies the user as C1-1-1. On the other hand, if the average tone value is equal to or less than the second threshold T2 (step S3; No), the classification unit 4 classifies the user as C1-1-2. The second threshold T2 is also determined by the above-mentioned analysis.

[0049] (4) If the user is determined to be C1-2 (step S2; Yes), the classification unit 4 determines whether the absolute value of the amount of change in tone value from morning to afternoon exceeds the third threshold T3 (step S4). If the classification unit 4 determines that the absolute value of the amount of change in tone value from morning to afternoon exceeds the third threshold T3 (step S4; Yes), it classifies the user as C1-2-1. On the other hand, if the classification unit 4 determines that the absolute value of the amount of change in tone value from morning to afternoon is equal to or less than the third threshold T3 (step S4; No), it classifies the user as C1-2-2. This third threshold T3 is also determined based on the above-mentioned analysis results.

[0050] As described above in detail, the classification device 1 according to this embodiment performs cluster analysis by having a plurality of subjects select a converted image CP that is closest to how the subject in the original image OP appears when actually viewed, and classifies users based on their selection results according to the classification criterion R generated by the analysis, so that the visual characteristics of the users can be identified and the identified visual characteristics can be utilized. A display device and a printing device that utilize the identified visual characteristics of the users will be described later.

[0051] In this embodiment, the user selects a converted image CP that is similar to the way the user sees the image from among a plurality of converted images CP, thereby classifying the user. However, there are cases where the user is unsure of which converted image CP to select. To deal with such cases, the configuration of the classification device 1 may be as shown in FIG. 10.

[0052] The classification device 1 shown in Fig. 10 is the same as the classification device 1 shown in Fig. 1 in that it includes a storage unit 2, a selection unit 3, and a classification unit 4. The classification device 1 shown in Fig. 10 differs from the classification device 1 shown in Fig. 1 in that the storage unit 2 stores a language database W. The classification device 1 shown in Fig. 10 can also classify users by selecting a converted image CP, as with the classification device 1 shown in Fig. 1, but in addition to or instead of the classification by selecting the converted image CP, it has a function of classifying users by having the user answer questions about their lifestyle habits, how they feel about vision at a certain time period, or their visual preferences.

[0053] In order to generate this classification device 1, a questionnaire survey is conducted in advance for multiple subjects who have undergone cluster analysis. This questionnaire survey collects answers to questions about the subjects' lifestyle habits, feelings about vision in a certain time period, or preferences about vision. By comparing these answers with the clusters generated by the cluster analysis, the correspondence between the multiple clusters generated by the cluster analysis and information expressing lifestyle habits, feelings about vision in a certain time period, or preferences about vision becomes clear. For example, as shown in FIG. 10, the language database W registers clusters generated as a result of cluster analysis of multiple subjects performed by the cluster analysis device 7. Furthermore, the language database W registers one of the multiple clusters in association with information expressing lifestyle habits, vision in the morning (feeling about vision in a certain time period), or preferences about vision. For example, a person whose lifestyle is "morning type" and whose vision in the morning is "high contrast" is associated with C1-2. Also, a person whose lifestyle is "night type" and whose vision in the morning is "glaring" is associated with C1-1-1. Visual preferences, such as "preferring light tones" and "preferring dark colors," are also associated with each cluster. In this way, the storage unit 2 stores a plurality of pieces of information representing lifestyle habits, visual sensations at certain times of the day, or visual preferences corresponding to the clusters.

[0054] The selection unit 3 allows the user to select information matching the user's lifestyle, visual perception in a certain time period, or visual preference from a plurality of pieces of information registered in association with a plurality of clusters in the language database W of the storage unit 2. For example, the selection unit 3 displays sentences expressing the lifestyle, visual perception in a certain time period, or visual preference registered in the language database W, and allows the user to select and input whether the sentence applies to the user.

[0055] The classification unit 4 classifies users based on information indicating lifestyle habits, perception of visual matters at a certain time period, or visual preferences selected by the user in the selection unit 3. For example, when the sentence "It is in the morning" is selected, the classification unit 4 classifies the user who made the selection into C1-2 (C1-2-1 or C1-2-2) by referring to the language database W. In addition, when the sentence "Visibility is dazzling in the morning" is selected, the classification unit 4 classifies the user who made the selection into C1-1-1.

[0056] It should be noted that the sense of vision in a certain time period is not limited to "how it looks in the morning." It may be during the day or at night. Also, instead of a method in which the user selects a sentence, a method in which the user selects information by answering Yes / No to a question may be adopted.

[0057] <Embodiment 2> Next, a description will be given of a second embodiment of the present invention. In the second embodiment, a display device 5 according to the first embodiment uses the classification results of the classification device 1, as shown in FIG.

[0058] The display device 5 includes the classification device 1, a storage unit 30, a readout unit 31, an image quality conversion unit 32, and a display unit 33. The hardware configuration of the display device 5 is the same as the hardware configuration of the classification device 1 according to the above-mentioned embodiment 1 shown in Fig. 7. The external memory 12 of the display device 5 stores a database DB described later.

[0059] The configuration and operation of the classification device 1 are as described in the above embodiment 1. In the classification device 1, the classification result of a user is input to the reading unit 31. For example, one of the clusters C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2 to which the user belongs is input to the reading unit 31.

[0060] The storage unit 30 stores a database DB in which clusters C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2 are associated with a morning tone curve TC and an afternoon tone curve TC. In this database DB, for example, tone curves TC-4 and TC-2 are associated with cluster C1-1-1 as the morning and afternoon tone curves TC. This association is based on the above-mentioned analysis results (see FIGS. 4 and 5). In FIG. 5, cluster C1-1-1 is associated with tone curves TC-4 and TC-2 because the average of the morning tone values ​​is close to 4 and the average of the afternoon tone values ​​is close to 2.

[0061] In the above example, the closest tone curve is selected, but to get closer to the average tone value, it is easy to add an interpolation process using images of two types of tone curves. For example, if the tone value is 3.5, it can be approximated by taking the brightness average of the TC-3 and TC-4 images (it is fine to use the average value on the tone curve, not after the actual image generation).

[0062] The reading unit 31 acquires from the timer 17 for the current time period and selects and reads out from the multiple tone curves TC-1 to TC-5 stored in the storage unit 30 a tone curve TC according to the visual characteristics of the user in the current time period corresponding to the cluster of the user classified by the classification device 1. Here, if the user is classified into cluster C1-1-1 and the current time period is morning, the reading unit 31 selects tone curve TC-1 and reads out tone curve TC-1 from the storage unit 30.

[0063] The image quality conversion unit 32 converts the image quality of the image to be displayed using the tone curve TC read by the reading unit 31. If the read tone curve TC is tone curve TC-1, the image quality conversion unit 32 performs image quality conversion using the tone curve TC-1. The display unit 33 displays the image converted by the image quality conversion unit 32.

[0064] In this way, the display device 5 according to the present embodiment can display images with image quality that matches the visual characteristics of the individual and does not cause discomfort. Even if there is a difference in how things look in the morning and afternoon, it is possible to select tone curves TC-1 to TC-5 that match the time of day and display images with image quality that does not cause discomfort in either the morning or afternoon.

[0065] It is also possible to use such a display device 5 for a remote conference. As shown in Fig. 12, the two devices are connected via a communication network while being placed in environments with a time difference. In this way, a two-way remote conference can be held using the display device 5 (camera 18A and microphone 18B).

[0066] For example, assume that the current time at the installation location of one display device 5 is in the morning, and the current time at the installation location of the other display device 5 is in the afternoon. In this case, an image captured by the camera 18A of one display device 5 is transmitted to the other display device 5 via a communication network and displayed. Here, since the image captured by the camera 18A is an image captured in the morning, if it is displayed as is, the image displayed on the display 14 of the other display device 5 viewed in the afternoon may be an unnatural image. However, according to this embodiment, the image quality is adjusted by a tone curve TC according to the appearance in the afternoon in the other display device 5, so that an image that does not feel unnatural according to the time zone of the other display device 5 can be displayed.

[0067] The display device 5 may have a configuration as shown in Fig. 13. This display device 5 includes a storage unit, a selection unit 3, an image quality conversion unit 32, and a display unit 33. The storage unit 2 stores an original image OP, a converted image CP, and a tone curve TC.

[0068] The selection unit 3 allows the user to select, for each expected time period, a converted image CP that is closest to how the subject will look when viewed during the expected time period when it is expected that the subject will actually be viewed, from among multiple converted images CP generated by converting the image quality of the same original image OP using multiple tone curves TC. This selection by the selection unit 3 is as described in the above embodiment 1. For example, if the expected time periods are morning and afternoon, the converted image CP for the morning and the converted image CP for the afternoon are selected.

[0069] The image quality conversion unit 32 converts the image quality of the image to be displayed with the tone curve TC used to generate the converted image CP selected for the expected time period corresponding to the current time period, from among the multiple tone curves TC, with reference to the timer 17. For example, if the current time period is morning, the tone curve TC corresponding to the converted image CP selected in the morning is read from the storage unit 2 and used to convert the image quality of the image. Even in this way, it is possible to display an image with image quality that matches the visual characteristics of the user on the display unit 33.

[0070] <Embodiment 3> Next, a third embodiment of the present invention will be described. In this embodiment, as shown in FIG. 14, a printing device 6 including the classification device 1 according to the first embodiment will be described.

[0071] The printing device 6 includes the classification device 1, a storage unit 30, a reading unit 35, an image quality conversion unit 36, and a printing unit 37. The hardware configuration of the printing device 6 is the same as the hardware configuration of the classification device 1 according to the above-mentioned first embodiment shown in Fig. 7. The configuration and operation of the classification device 1 are as described in the above-mentioned first embodiment.

[0072] In the classification device 1, the classification result of the user is input to the readout unit 35. For example, one of the clusters C1-1-1, C1-1-2, C1-2-1, C1-2-2, and C2 to which the user belongs is input to the readout unit 35. The storage unit 30 stores a database DB, a tone curve TC for image display, and a tone curve TC for printing. The tone curves for printing include a standard tone curve, a morning tone curve TC-a1, and an afternoon tone curve TC-a2. The standard tone curve is a general-purpose tone curve that converts an image to be printed and an image displayed on the display 14 to the same image quality as much as possible. The morning tone curve TC-a1 is a tone curve that converts an image to an image quality close to how the subject actually looks when viewed in the morning. The afternoon tone curve TC-a2 is a tone curve that converts an image to an image quality close to how the subject actually looks when viewed in the afternoon. As shown in Figure 15, in an overall comparison, tone curve TC-a2 is located higher on the output axis than tone curve TC-a1, resulting in a conversion that is relatively darker in the morning and brighter in the afternoon. Also, looking at each tone curve in part, tone curve TC-a1 is specified so that the contrast in areas with high luminance is relatively large, while tone curve TC-a2 is specified so that the contrast in areas with low luminance is relatively large.

[0073] Returning to FIG. 14, the readout unit 35 first reads out the tone curve for the current time period of the cluster to which the user belongs from among the tone curves TC-1 to TC-5. Furthermore, the readout unit 35 reads out the printing tone curve and the printing tone curves TC-a1 or TC-a2 corresponding to the current time period, morning, and afternoon from the storage unit 30. The image quality conversion unit 36 ​​converts the image quality of the image to be displayed using one of the image tone curves TC1 to TC5 read out by the readout unit 35 to generate a converted image CP. At this stage, the converted image has image quality that conforms to the user's personal visual characteristics.

[0074] Furthermore, the image quality conversion unit 36 ​​converts the image quality of the converted image CP with the tone curve (TC-a1 or TC-a2) for the current time period (morning or afternoon) so that the converted image CP will have an appropriate tone when converted with the standard tones for printing. In this way, the image printed on the printing paper can be corrected in advance so that the image is converted to an image quality close to how the subject actually appears when viewed in the current time period.

[0075] Furthermore, the image quality conversion unit 36 ​​further converts the image quality of the converted image CP using a standard tone curve for printing. The printing unit 37 prints this converted image CP using the printer 22 (see FIG. 7). At this stage, the image quality of the converted image CP displayed on the display 14 matches that of the printed image. In this way, the image appears as similar as possible in image quality on the printed material, which is a reflective object, and on the display screen, which is a luminous object.

[0076] The final printed image matches the visual characteristics of the user, has the same image quality as the image displayed on the display 14, and is an image with image quality that matches the time of day, morning or afternoon.

[0077] In the above embodiments, the time period is divided into morning and afternoon, and converted images CP that are close to the actual appearance of the subject in the morning and afternoon are selected. However, this is not limited to this. The time period may be further divided, and converted images CP that are close to the actual appearance of the subject at each time may be selected. Conversely, the time period may not be distinguished, and the user may select one converted image CP and apply the tone curve TC.

[0078] Also, images may be prepared for each region where the user lives, and the user may select a converted image CP having as its subject the scenery of the region where the user lives. If the user is familiar with that scenery, he or she will usually recognize how the sun hits the subject at that time of day, and therefore, when selecting a converted image CP, the user will be less likely to feel uncomfortable with the image.

[0079] In the above embodiment, the user is prompted to select converted images CP of different subjects and different time periods. In this way, a tone curve TC that more accurately matches the visual characteristics can be selected. However, the converted images CP may be only those images of the same subject whose image quality has been converted, or only one converted image CP may be selected regardless of the time period.

[0080] Furthermore, in the above embodiment, a plurality of subjects are classified by hierarchical cluster analysis, but a plurality of subjects may be classified by non-hierarchical cluster analysis.

[0081] Furthermore, the hardware configuration and software configuration of the classification device 1 and the like are merely examples and can be changed and modified as desired.

[0082] The core part of the processing of the classification device 1, etc., which is composed of the storage unit 2, the selection unit 3, the classification unit 4, etc., can be realized by using a normal computer system, not a dedicated system. For example, the classification device 1, etc. that executes the above processing may be configured by storing and distributing a computer-readable recording medium (flexible disk, CD-ROM, DVD-ROM, etc.) for executing the above operations, and installing the computer program on a computer. Also, the classification device 1, etc. may be configured by storing the computer program in a storage device of a server device on a communication network such as the Internet, and downloading it by a normal computer system.

[0083] When the functions of the classification device 1 and the like are realized by sharing between an OS (operating system) and an application program, or by cooperation between the OS and the application program, only the application program portion may be stored in a recording medium or storage device.

[0084] It is also possible to superimpose a computer program on a carrier wave and distribute it via a communication network. For example, the computer program may be posted on a bulletin board system (BBS) on the communication network and distributed via the network. Then, the computer program may be started and executed under the control of the OS in the same way as other application programs, thereby enabling the above-mentioned processing to be performed.

[0085] Various embodiments and modifications of the present invention are possible without departing from the broad spirit and scope of the present invention. The above-described embodiments are for the purpose of explaining the present invention and do not limit the scope of the present invention. That is, the scope of the present invention is indicated by the claims, not the embodiments. Various modifications made within the scope of the claims and the scope of the invention equivalent thereto are considered to be within the scope of the present invention. [Industrial Applicability]

[0086] The present invention can be applied to a wide range of fields, such as image display and printing. [Explanation of symbols]

[0087] 1 Classification device, 2 Memory unit, 3 Selection unit, 4 Classification unit, 5 Display device, 6 Printing device, 7 Cluster analysis device, 10 CPU, 11 Memory, 12 External memory, 13 Operation unit, 14 Display, 15 Communication interface, 16 Input / output unit, 17 Timer, 18A Camera, 18B Microphone, 19 Program, 20 Internal bus, 21 Recording medium, 22 Printer, 30 Memory unit, 31 Readout unit, 32 Image quality conversion unit, 33 Display unit, 35 Readout unit, 36 Image quality conversion unit, 37 Printing unit, OP, A1-A3, P1-P3 Original image, CP Converted image, TC, TC-1-TC-5 Tone curve, DB Database, W Language database, R Classification criteria

Claims

1. a selection unit that allows a user to select a converted image that is closest to how a subject in an original image looks when actually viewed from among a plurality of converted images obtained by converting the image quality of an original image using a plurality of tone curves having different brightness and contrast conversion characteristics; a classification unit that classifies users based on the selection results of the converted images by the users in the selection unit according to classification criteria generated by cluster analysis using the selection results of the converted images by a plurality of subjects; Equipped with In the cluster analysis, the classification criteria are generated by classifying a plurality of subjects in a vector space in which a coordinate axis indicating a numerical value corresponding to a selection result of a converted image that is closest to how the subject looks when actually viewed is independently defined for each expected time period when the subject is expected to be viewed; The selection unit is having a user select a converted image that is closest to how the subject appears when actually viewed for each of the estimated time periods, and determining a vector that indicates the user's selection result in the vector space; The classification unit is classifying users based on vectors indicating user selections in the vector space; Classification device.

2. A classification device according to claim 1 ; an image quality conversion unit that converts the image quality of a displayed image using a tone curve corresponding to the user's cluster classified by the classification device, out of a plurality of tone curves having different brightness and contrast conversion characteristics; A display device comprising:

3. A classification method performed by a classification device, comprising: a cluster analysis is performed using the results of selecting a converted image that is closest to how the subject in the original image looks when actually viewed from a plurality of converted images obtained by converting the image quality of the original image using a plurality of tone curves each having different brightness and contrast conversion characteristics, for each estimated time period during which the plurality of subjects are expected to view the subject; and a classification criterion is generated for classifying the plurality of subjects in a vector space in which coordinate axes indicating values ​​corresponding to the selection results of the converted image that is closest to how the subject looks when actually viewed are independently defined for each estimated time period; having a user select, for each of the assumed time periods, a converted image that is closest to how the subject in the original image appears when actually viewed from among the plurality of converted images, and determining a vector in the vector space that indicates the user's selection result; classifying the user based on the user's selection of the transformed image in the vector space according to the classification criteria; Classification method.

Images